-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathREADME.txt
51 lines (39 loc) · 2.91 KB
/
README.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
Steps for running DisCoveR-py within the docker container:
- Open the CLI (Command Line Interface) of the Docker Container discover-py
- Run "python main.py -f -a" to replicate the experiments of the paper (depending on your machine it might be an overnight job)
- In powershell/bash do 'docker ps' and copy the 'CONTAINER ID' of the container with the name "discover-py"
- In poweshell/bash do 'docker cp {CONTAINER ID}:/DisCoveR-py/models .' to copy all the results from the containers models folder in the current host folder (replace the '.'(dot) in the command to use a specific folder)
-The results are suffixed as follows:
- "{dataset name}_model" files contain the DCR graphs
- "{dataset name}_timings" folders contain image files ".jpg" of the timing data:
- "{CONDITION|RESPONSE}_{event from}_{event to}_hist" for histograms
- "{CONDITION|RESPONSE}_{event from}_{event to}_boxplot" for boxplots
- "{CONDITION|RESPONSE}_{event from}_{event to}_simple_fit" for the best 5 single parametric distribution fit from the Fitter library
- "{CONDITION|RESPONSE}_{event from}_{event to}_advanced_fit" (only applicable for the subset of the road traffic fine dataset mined conditions and responses that have advanced fitting initial parameters specified in the advanced_timings_fit method of the main.py file)
Experiment setup:
The code was run on Linux Ubuntu 20.4 OS inside a Windows 10 Subsystem for Linux WSL2 installation with the following specs:
Processor: Intel(R) Core(TM) i7-7660U CPU @ 2.50GHz 2.50 GHz
RAM: 16GB
64-bit OS, x64-based processor
Intel integrated graphics
The expected runtime is around 8 hours.
Manual Steps (not necessary unless something goes wrong):
Prerequisites:
Make sure that the project folder contains the following directory structure:
DisCoveR-py
data
discover
models
main.py
Using Docker run the Dockerfile
- (done in the Dockerfile) Place the event log files (.xes) downloaded and unarchived from their ".gz" format from the following links inside the "data" folder:
- https://data.4tu.nl/ndownloader/files/24073733
- https://data.4tu.nl/ndownloader/files/24027287
- https://data.4tu.nl/ndownloader/files/24063818
- https://data.4tu.nl/ndownloader/files/24018146
- Run main.py with arguments:
- '-f' or '--fine' for creating the advanced timing distributions fit from the road traffic fine dataset
- '-a' or '--all' for creating the summary statistics for all the 4 event logs
- Make sure the Docker container is running. In powershell/bash do
- 'docker ps' and copy the 'CONTAINER ID' of the container with the name "discover-py"
- 'docker cp {CONTAINER ID}:/DisCoveR-py/models .' to retrieve all the results from the models folder in the current folder (replace the '.'(dot) in the command to use a specific folder)